Long-term seizure outcome in patients initially seizure-free after resective epilepsy surgery.

OBJECTIVE: To evaluate the likelihood of and risk factors for seizure recurrence in patients initially seizure-free after resective surgery for intractable epilepsy. METHODS: One hundred seventy-five patients who underwent lobectomy between 1972 and 1992 and were seizure-free during the first postoperative year were retrospectively studied. Outcome was measured by relapse risk, presence of auras in otherwise seizure-free patients, and seizure frequency among relapsers. Factors significant in bivariate or Kaplan-Meier analysis or considered potentially predictive a priori were included in multivariate models. RESULTS: Of the 175 patients (mean follow-up 8.4 years), 63% never relapsed. The likelihood of being seizure-free was 83 +/- 6% 3 years after surgery, 72 +/- 7% after 5 years, and 56 +/- 9% after 10 years. After adjusting for age at surgery, duration of preoperative epilepsy, and resection site, normal pathology was associated with increased risk of relapse compared to mesial temporal sclerosis or other pathology (p = 0.036; hazard ratio [HR] 2.38; 95% CI 1.06 to 5.34). Among patients otherwise seizure-free, preoperative illness of > or =20 years was associated with increased risk of postoperative auras (p = 0.040; HR 3.55; 95% CI 1.06 to 11.90). Among relapsers, 51% experienced one or fewer seizures per year. Normal pathology and earlier relapse were associated with higher postoperative seizure frequency. CONCLUSIONS: In patients seizure-free during the first year after resective epilepsy surgery, the likelihood of remaining seizure-free declined to 56% over 10 years, but half of patients who relapsed had at most one seizure per year. Longer preoperative illness and normal pathology predicted poorer outcome.